Search

Abstract Unmanned aerial vehicle (UAV)-assisted communication based on automatic modulation classification (AMC) technology is considered an effective solution to improve the transmission efficiency of wireless communication systems, as it can adaptively select the most suitable modulation method according to the current communication environment. However, many existing deep learning (DL)-based AMC methods cannot be directly applied to UAV platform with limited computing power and storage space, because of the contradiction between accuracy and efficiency. This paper mainly studies the lightweight of DL-based AMC networks to improve adaptability in resource-constrained scenarios. To address this challenge, we propose an ultra-lightweight neural network (ULNN). This network incorporates a lightweight convolutional structure, attention mechanism, and cross-channel feature fusion technique. Additionally, we introduce data augmentation (DA) based on signal phase offsets during the model training process, aimed at improving the model’s generalization ability and preventing overfitting. Through experimental validation on the public dataset RML2016.10 A, the ULNN we proposed achieves an average precision of 62.83% with only 8815 parameters and reaches a peak classification accuracy of 92.11% at SNR = 10 dB. The experimental results show that ULNN can achieve high recognition accuracy while keeping the model lightweight, and is suitable for UAV platform with limited resources.

Abstract Automatic Modulation Classification (AMC) is crucial in non-cooperative communication systems as it facilitates the identification of interference signals with minimal prior knowledge. Although there have been significant advancements in Deep Learning (DL) within the field of AMC, leveraging the inherent relationships between In-phase (I) and Quadrature-phase (Q) components, and enhance recognition accuracy under low signal-to-noise ratio (SNR) conditions remains a challenge. This study introduces a complex-valued convolutional fusion-type multi-stream spatiotemporal network (CC-MSNet) for AMC, which combines spatial and temporal feature extraction modules for modulation recognition. Experimental results demonstrate that the CC-MSNet performs well on three benchmark datasets, RML2016.10a, RML2016.10b, and RML2016.04c, with average recognition accuracy of 62.86%, 65.08%, and 71.12%. It also performs excellently in low SNR environments below 0dB, significantly outperforming other networks.

Abstract Fixed frequency beam-scanning leaky-wave antennas have been a focus of attention for many scholars in recent years, and numerous related results have been obtained. However, these antennas suffer from several issues such as small beam-scanning range, low gain, and unsatisfactory impedance matching. To address these problems, this paper proposes a microstrip line (ML) antenna unit based on liquid crystal (LC) materials etched Complementary Split Ring Resonator (CSRR). In a first-of-its-kind approach, the substrate integrated waveguide (SIW) structure and the ML transmission structure are combined to present the SIW-ML transmission structure. The antenna operates in the Ka-band with excellent resonance characteristics at 34.7 GHz, and the S11 parameters are below − 13 dB in the frequency range of 30–40 GHz, indicating outstanding impedance matching. By arranging 56 antenna units, a periodic leaky-wave antenna is created, enabling fixed-frequency beam-scanning at 34.7 GHz. Experimental results show that the antenna can achieve scanning of angles between − 53° and + 60° with a gain of up to 12.63 dB. Once single-beam scanning is achieved, a method combining LC and discrete amplitude weighting technique, as well as multi-beam theory, is proposed for multi-beam study. Experimental results reveal that the designed 56-unit beam-scanning antenna can effectively realize beam scanning in two directions.

Automatic Modulation Recognition (AMR) is a key technology in the field of cognitive communication, playing a core role in many applications, especially in wireless security issues. Currently, deep learning (DL)-based AMR technology has achieved many research results, greatly promoting the development of AMR technology. However, the few-shot dilemma faced by DL-based AMR methods greatly limits their application in practical scenarios. Therefore, this paper endeavored to address the challenge of AMR with limited data and proposed a novel meta-learning method, the Multi-Level Comparison Relation Network with Class Reconstruction (MCRN-CR). Firstly, the method designs a structure of a multi-level comparison relation network, which involves embedding functions to output their feature maps hierarchically, comprehensively calculating the relation scores between query samples and support samples to determine the modulation category. Secondly, the embedding function integrates a reconstruction module, leveraging an autoencoder for support sample reconstruction, wherein the encoder serves dual purposes as the embedding mechanism. The training regimen incorporates a meta-learning paradigm, harmoniously combining classification and reconstruction losses to refine the model’s performance. The experimental results on the RadioML2018 dataset show that our designed method can greatly alleviate the small sample problem in AMR and is superior to existing methods.

Abstract Automatic modulation recognition (AMR) is a critical technology in spatial cognitive radio (SCR), and building high-performance AMR model can achieve high classification accuracy of signals. AMR is a classification problem essentially, and deep learning has achieved excellent performance in various classification tasks. In recent years, joint recognition of multiple networks has become increasingly popular. In complex wireless environments, there are multiple signal types and diversity of characteristics between different signals. Also, the existence of multiple interference in wireless environment makes the signal characteristics more complex. It is difficult for a single network to accurately extract the unique features of all signals and achieve accurate classification. So, this article proposes a time–frequency domain joint recognition model that combines two deep learning networks (DLNs), to achieve higher accuracy AMR. A DLN named MCLDNN (multi-channel convolutional long short-term deep neural network) is trained on samples composed of in-phase and quadrature component (IQ) signals, to distinguish modulation modes that are relatively easy to identify. This paper proposes a BiGRU3 (three-layer bidirectional gated recurrent unit) network based on FFT as the second DLN. For signals with significant similarity in the time domain and significant differences in the frequency domain that are difficult to distinguish by the former DLN, such as AM-DSB and WBFM, FFT (Fast Fourier Transform) is used to obtain frequency domain amplitude and phase (FDAP) information. Experiments have shown that the BiGUR3 network has superior extraction performance for amplitude spectrum and phase spectrum features. Experiments are conducted on two publicly available datasets, the RML2016.10a and RML2016.10b, and the results show that the overall recognition accuracy of the proposed joint model reaches 94.94% and 96.69%, respectively. Compared to a single network, the recognition accuracy is significantly improved. At the same time, the recognition accuracy of AM-DSB and WBFM signals has been improved by 17% and 18.2%, respectively.

This article proposes a novel fixed-frequency beam scanning leakage antenna based on a liquid crystal metamaterial (LCM) and adopting a metal column embedded microstrip line (MCML) transmission structure. Based on the microstrip line (ML) transmission structure, it was observed that by adding two rows of metal columns in the dielectric substrate, electromagnetic waves can be more effectively transmitted to reduce dissipation, and attenuation loss can be lowered to improve energy radiation efficiency. This antenna couples TEM mode electromagnetic waves into free space by periodically arranging 72 complementary split ring resonators (CSRRs). The LC layer is encapsulated in the transmission medium between the ML and the metal grounding plate. The simulation results show that the antenna can achieve a 106° continuous beam turning from reverse −52° to forward 54° at a frequency of 38 GHz with the holographic principle. In practical applications, beam scanning is achieved by applying a DC bias voltage to the LC layer to adjust the LC dielectric constant. We designed a sector-blocking bias feeder structure to minimize the impact of RF signals on the DC source and avoid the effect of DC bias on antenna radiation. Further comparative experiments revealed that the bias feeder can significantly diminish the influence between the two sources, thereby reducing the impact of bias voltage introduced by LC layer feeding on antenna performance. Compared with existing approaches, the antenna array simultaneously combines the advantages of high frequency band, high gain, wide beam scanning range, and low loss.

In recent years, the three-dimensional (3D) radar detection range has played an essential role in the layout of devices such as aircraft and drones. To compensate for the shortcomings of three-dimensional calculations for radar terrain masking, a new calculation method is proposed for assessing the terrain occlusion of radar detection range. First, the high-dimensional electromagnetic data after discretization are modeled based on the tensor data structure, and the tensor grid dilation operator is constructed. Then, the dilation process begins from the overlapping section of the radar detection range and terrain, and it is adjusted by the terrain occlusion judgment factor and the dilation judgment factor to obtain the obstructed part due to the terrain. Finally, the actual radar detection range under terrain occlusion is obtained. The simulation results show that the method proposed in this paper can adapt to different grid sizes and terrain shapes, significantly enhancing computational efficiency while maintaining internal features.

The graph neural network (GNN) has shown outstanding performance in processing unstructured data. However, the downstream task performance of GNN strongly depends on the accuracy of data graph structural features and, as a type of deep learning (DL) model, the size of the training dataset is equally crucial to its performance. This paper is based on graph neural networks to predict and complete the target radio environment map (REM) through multiple complete REMs and sparse spectrum monitoring data in the target domain. Due to the complexity of radio wave propagation in space, it is difficult to accurately and explicitly construct the spatial graph structure of the spectral data. In response to the two above issues, we propose a multi-source domain adaptive of GNN for regression (GNN-MDAR) model, which includes two key modules: (1) graph structure alignment modules are used to capture and learn graph structure information shared by cross-domain radio propagation and extract reliable graph structure information for downstream reference signal receiving power (RSRP) prediction task; and (2) a spatial distribution matching module is used to reduce the feature distribution mismatch across spatial grids and improve the model’s ability to remain domain invariant. Based on the measured REMs dataset, the comparative results of simulation experiments show that the GNN-MDAR outperforms the other four benchmark methods in accuracy when there is less RSRP label data in the target domain.

Understanding the yield attributes of the popular rice (Oryza sativa L.) hybrids can provide useful information for developing new hybrid rice varieties. Field experiments were conducted at two subtropical sites (Hengyang and Xingyi) in two years (2021 and 2022) to compare grain yield and yield attributes of three currently popular hybrid rice varieties (Jingliangyouhuazhan, Jingliangyou 534, and Longliangyouhuazhan) with three representative super hybrid rice varieties (Y-liangyou 1, Y-liangyou 2, and Y-liangyou 900). No significant differences in grain yield were observed between the group of popular hybrids and the group of super hybrids at Hengyang and Xingyi in 2021 or at Xingyi in 2022, but at Hengyang in 2022, the group of popular hybrids produced a 15% higher grain yield. The grain yield at Hengyang in 2022 was lower than that at Hengyang and Xingyi in 2021 and at Xingyi in 2022. At Hengyang in 2022, the group of popular hybrids had 9–15% higher panicles per m2, spikelet filling percentage, and harvest index but similar spikelets per panicle and total biomass production and 12% lower grain weight compared to the group of super hybrids. Correlation plot analysis showed that grain yield was significantly related to panicles per m2 but not other yield attributes across six varieties at Hengyang in 2022. These results indicated that the currently popular hybrid rice varieties had higher yield performance than the representative super hybrid rice varieties under the condition of relatively lower productivity, and the key factor for this higher yield performance in the popular hybrids was the higher panicles per m2. This finding highlights that more attention should focus on the yield performance under medium- and low-productivity conditions in hybrid rice breeding programs, and an effective breeding strategy is to select the varieties with high panicle numbers.

Improving grain quality is critical to the development of hybrid rice. In recent years, multiple hybrid rice cultivars with superior grain quality have been developed in China through dedicated breeding efforts. However, limited information is available on yield performance and yield formation characteristics of these high-quality hybrid rice cultivars. In this study, field experiments were conducted to compare yield attributes between a recently developed indica hybrid rice with superior grain quality (Jingliangyou 1468, JLY1468) and a relatively older indica hybrid rice cultivar with high yield potential (Liangyoupeijiu, LYPJ). Results showed that grain yield was 13–17% higher in JLY1468 than in LYPJ. The higher grain yield of JLY1468 than of LYPJ was mainly attributable to a higher spikelet filling percentage, which was closely related to a higher harvest index resulting from a higher crop growth rate during the post-heading period. The results of this study suggest that late-stage vigor may be a vital trait for the development of indica hybrid rice cultivars with both high quality and high yield.

In general, judging the use/idle state of the wireless spectrum is the foundation for cognitive radio users (secondary users, SUs) to access limited spectrum resources efficiently. Rich information can be mined by the inherent correlation of electromagnetic spectrum data from SUs in time, frequency, space, and other dimensions. Therefore, how to efficiently use the spectrum status of each SU implementation of reception multidimensional combination forecasting is the core of this paper. In this paper, we propose a deep-learning hybrid model called TensorGCN-LSTM based on the tensor data structure. The model treats SUs deployed at different spatial locations under the same frequency, and the spectrum status of SUs themselves under different frequencies in the task area as nodes and constructs two types of graph structures. Graph convolutional operations are used to sequentially extract corresponding spatial-domain and frequency-domain features from the two types of graph structures. Then, the long short-term memory (LSTM) model is used to fuse the spatial, frequency, and temporal features of the cognitive radio environment data. Finally, the prediction task of the spectrum distribution situation is accomplished through fully connected layers. Specifically, the model constructs a tensor graph based on the spatial similarity of SUs’ locations and the frequency correlation between different frequency signals received by SUs, which describes the electromagnetic wave’s dependency relationship in spatial and frequency domains. LSTM is used to capture the electromagnetic wave’s dependency relationship in the temporal domain. To evaluate the effectiveness of the model, we conducted ablation experiments on LSTM, GCN, GC-LSTM, and TensorGCN-LSTM models using simulated data. The experimental results showed that our model achieves better prediction performance in RMSE, and the correlation coefficient R2 of 0.8753 also confirms the feasibility of the model.

The demand for and consumption of soft-textured, high-quality rice with low amylose content have substantially increased in China with improved economic and living standards. Shifting sowing date is a common strategy to optimize temperature conditions and improve rice quality. In this study, a high-quality rice cultivar with low amylose content (Jingliangyou 1468) was grown under four sowing dates (May 15th and 25th, and June 4th and 14th) in 2019 and 2020 to determine the effects of sowing date on palatability-related traits and evaluate relationships between palatability-related traits and daily mean temperature during the grain-filling period. Results showed that delaying sowing date decreased daily mean temperature during the grain-filling period and consequently increased amylose content and setback viscosity of milled rice flour, which in turn increased hardness and chewiness of cooked rice. The finding of this study suggests that the strategy of delayed sowing does not improve the palatability of high-quality rice in the context of consumers’ preference for soft-textured rice.

Field experiments were conducted to compare two hybrid rice cultivars—a recently released high-quality cultivar (Jingliangyou 1468, JLY1468) and a relatively older cultivar (Liangyoupeijiu, LYPJ). Results showed that hardness, springiness, cohesiveness, resilience, and chewiness of cooked milled rice were all lower in JLY1468 than in LYPJ, due to its lower amylose content and altered paste properties of milled rice flour. Active digestion duration of cooked milled rice was 26% shorter and the glucose production rate from starch digestion was 33% faster in JLY1468 compared with LYPJ. Texture and starch digestion properties of cooked milled rice as a factor of temperature during the grain-filling period were different between LYPJ and JLY1468 due to differing amylose contents and gel consistencies of milled rice flour in response to temperature. This study highlights that attention should be paid to potential health risks associated with the development of high-quality hybrid rice cultivars with soft texture.

Head milled rice is the main form of rice for sale and consumption. However, previous studies on the yield change due to the development of new cultivars in rice generally focus on grain yield but few on head milled rice yield. In this study, field experiments were conducted in two years (2019 and 2020) to compare head milled rice yield and associated traits (grain yield, milled recovery traits, and shape and chalkiness traits of rice grains) between two middle-season hybrid rice cultivars released 18 years apart, i.e., Jingliangyou 1468 (JLY1468), a recently-released cultivar with high eating quality, and Liangyoupeijiu (LYPJ), an old cultivar with high grain yield. JLY1468 had higher head milled rice yield than LYPJ by 30% in 2019 and by 33% in 2020. The higher head milled rice yield in JLY1468 than in LYPJ was attributable to improvements in both grain yield, and head milled rice rate (HMRR). The improvement in HMRR in JLY1468 compared to LYPJ was mainly attributable to a reduction in chalkiness degree, which was associated with a decrease in rice grain size. The results of this study provide evidence for the improvement in head milled rice yield in middle-season hybrid rice with the development of new cultivars in recent years.

Radio signals are polluted by noise in the process of channel transmission, which will lead to signal distortion. Noise reduction of radio signals is an effective means to eliminate the impact of noise. Using deep learning (DL) to denoise signals can reduce the dependence on artificial domain knowledge, while traditional signal-processing-based denoising methods often require knowledge of the artificial domain. Aiming at the problem of noise reduction of radio communication signals, a radio communication signal denoising method based on the relativistic average generative adversarial networks (RaGAN) is proposed in this paper. This method combines the bidirectional long short-term memory (Bi-LSTM) model, which is good at processing time-series data with RaGAN, and uses the weighted loss function to construct a noise reduction model suitable for radio communication signals, which realizes the end-to-end denoising of radio signals. The experimental results show that, compared with the existing methods, the proposed algorithm has significantly improved the noise reduction effect. In the case of a low signal-to-noise ratio (SNR), the signal modulation recognition accuracy is improved by about 10% after noise reduction.

Rice noodles are usually manufactured using rice grains of the main crop. There is limited information available on rice noodles processed from ratoon rice grains. In this study, two-year field experiments were conducted to compare the cooking and texture properties of noodles and the grain chemical properties of ratoon crops with those of main crops from two rice cultivars (Guichao 2 and Zhenguiai) that are widely used for processing noodles. Results showed that the cooked break rate and the cooking loss rate of rice noodles processed from grains of the ratoon crop were similar to those of the main crop in both cultivars; however, changes in texture of cooked rice noodles processed from grains of the ratoon crop compared with the main crop were cultivar-dependent, being significantly softer in Zhenguiai but not in Guichao 2. Hardness and chewiness of cooked rice noodles were significantly negatively correlated with amylopectin content in milled rice grains in Zhenguiai, indicating that amylopectin content is a key chemical property determining the texture differences between rice noodles processed from grains of ratoon and main crops.

Radar signal anomaly detection is an effective method to detect potential threat targets. Given the low Accuracy of the traditional AE model and the complex network of GAN, an anomaly detection method based on ResNet-AE is proposed. In this method, CNN is used to extract features and learn the potential distribution law of data. LSTM is used to discover the time dependence of data. ResNet is used to alleviate the problem of gradient loss and improve the efficiency of the deep network. Firstly, the signal subsequence is extracted according to the pulse’s rising edge and falling edge. Then, the normal radar signal data are used for model training, and the mean square error distance is used to calculate the error between the reconstructed data and the original data. Finally, the adaptive threshold is used to determine the anomaly. Experimental results show that the recognition Accuracy of this method can reach more than 85%. Compared with AE, CNN-AE, LSTM-AE, LSTM-GAN, LSTM-based VAE-GAN, and other models, Accuracy is increased by more than 4%, and it is improved in Precision, Recall, F1-score, and AUC. Moreover, the model has a simple structure, strong stability, and certain universality. It has good performance under different SNRs.

Automatic modulation discrimination (AMC) is one of the critical technologies in spatial cognitive communication systems. Building a high-performance AMC model in intelligent receivers can help to realize adaptive signal synchronization and demodulation. However, tackling the intra-class diversity problem is challenging to AMC based on deep learning (DL), as 16QAM and 64QAM are not easily distinguished by DL networks. In order to overcome the problem, this paper proposes a joint AMC model that combines DL and expert features. In this model, the former builds a neural network that can extract the time series and phase features of in-phase and quadrature component (IQ) samples, which improves the feature extraction capability of the network in similar models; the latter achieves accurate classification of QAM signals by constructing effective feature parameters. Experimental results demonstrate that our proposed joint AMC model performs better than the benchmark networks. The classification accuracy is increased by 11.5% at a 10 dB signal-to-noise ratio (SNR). At the same time, it also improves the discrimination of QAM signals.